The centerpiece of the original evidence was a laboratory game nicknamed ‘Angry Blueberries,’ in which participants fired shots at targets to earn points. Some participants were randomly assigned a large budget of shots, making them ‘rich,’ while others received a small budget, making them ‘poor.’ The original researchers reported that ‘poor’ participants spent more time aiming and earned more points per shot, which they interpreted as a focusing benefit of scarcity: with fewer resources, people concentrate harder. A second manipulation allowed some participants to borrow shots from future levels at a punishing 100 percent interest rate. The ‘poor’ borrowed far more than the ‘rich’ and suffered for it, which the original authors read as the dark side of tunneling: scarcity makes people neglect information outside their immediate tunnel, such as the true cost of borrowing.

Tho Pesch and colleagues do not dispute that these behavioral differences exist; their re-analysis of both the original dataset and a large-scale 2019 self-replication by the same authors reproduced them faithfully. What they contest is the interpretation. The first problem concerns the celebrated performance advantage of the ‘poor.’ In the game, earlier shots were systematically more profitable than later ones, because more targets remained available at the start of each level. Participants scored an average of 2.81 points on their first shot of each level, dropping to 1.80 points by the third shot and a mere 0.69 points by the fifteenth. Because ‘poor’ participants could take only three shots per level, they were structurally barred from the low-yield later shots that ‘rich’ participants inevitably used, inflating the ‘poor’ participants’ average.

To test whether the ‘poor’ were genuinely more focused, the re-analysis compared the two groups only on the first three shots of each level, the shots both groups could take. If scarcity truly sharpened attention, the ‘poor’ should have outperformed the ‘rich’ even on this level playing field. They did not. In the original study there were no significant differences in points per shot across the first three shots, and the large-scale replication, which was well powered, showed no meaningful difference at all. The longer aiming times of the ‘poor’ did not translate into better outcomes, casting doubt on the claim that this behavior reflected heightened focus. It could just as plausibly reflect hesitation or inattention, the researchers note, and the original study was underpowered to distinguish between these possibilities.

The borrowing result fared no better under scrutiny. The original authors argued that ‘poor’ participants over-borrowed because tunneling made them neglect the high interest rate. But the re-analysis revealed a far simpler mechanism: borrowing was only possible after a participant exhausted their budget, and ‘poor’ participants, with only three shots per level, ran out of shots on 45 percent of levels, while ‘rich’ participants, with fifteen shots, exhausted their budget on just 2 percent. The ‘poor’ were simply offered the borrowing decision far more often. When the researchers examined only those cases where participants actually had the chance to borrow, the pattern reversed: ‘rich’ participants always borrowed, whereas ‘poor’ participants borrowed in only 69 percent of cases. If anything, the ‘poor’ were more cautious about high-interest borrowing than the ‘rich,’ the opposite of what attentional neglect predicts.

A further statistical anomaly deepened the skepticism. The original paper reported a correlation between ‘poor’ participants’ aiming times and the number of shots they borrowed, but only on the specific subset of levels where borrowing occurred. The re-analysis found that this correlation vanishes when any other subset of observations is considered, and it fails to replicate in the 2019 large-scale replication, suggesting it was a fragile, subset-dependent result rather than a robust signature of tunneling.

To settle the question experimentally, the team ran a conceptual replication that held the decision environment constant for both groups. Instead of borrowing, participants on the first ten levels could sacrifice shots to buy insurance against a future ‘drought’ that might destroy points from a separate pineapple field. Both ‘rich’ and ‘poor’ participants received the insurance offer equally often, and eight of the ten insurances had positive expected value while two were ‘scams’ whose costs exceeded their benefits. Scarcity theory predicts that the ‘poor,’ fixated on present needs, would neglect future-oriented insurance, while the ‘rich’ would selectively buy the worthwhile policies. The opposite emerged. There was no meaningful difference between groups in purchasing positive-value insurance, with equivalence tests confirming statistical equivalence, and the ‘poor’ were actually significantly less likely than the ‘rich’ to buy the scam policies. The study also replicated the original behavioral effects, including longer aiming times and higher points per shot for the ‘poor,’ while again showing no performance difference on the first three shots, demonstrating that the manipulation worked even though the tunneling mechanism did not.

Beyond dismantling a single study, the paper raises a deeper conceptual problem: nobody can say precisely what belongs inside the scarcity ‘tunnel.’ Shah and colleagues defined tunneling as attention shifted toward pressing expenses today and away from the future, which would restrict it to intertemporal decisions. Other reviews define it more loosely as focus on poverty-related issues that enhances resource efficiency, a scope vague enough that borrowing could plausibly fall inside the tunnel. Without a falsifiable boundary, the authors warn, any decision the poor make well can be attributed post hoc to greater focus, and any decision they make badly to neglect, rendering the theory untestable as stated.

The researchers also emphasize what the ‘Angry Blueberries’ experiment might still teach, if read differently. Rather than revealing a scarcity-induced mindset, the study may illustrate how scarcity changes the structure of the decision space itself. In the game, as in real life, the ‘poor’ confront the choice of borrowing at inflated rates far more often than the ‘rich,’ and that structural exposure, not any cognitive tunnel, drives their worse outcomes. The authors suggest this reframing points the field toward studying the environmental constraints that shape behavior under adversity, an aspect they argue has been largely neglected. They caution that both the original studies and their own replication relied on artificial laboratory paradigms and online MTurk samples, limiting generalizability to real-world poverty.

The findings arrive amid broader scrutiny of scarcity research, with recent audits questioning the replicability of several studies in the field. This new work makes a subtler point: even findings that replicate can mislead, if the interpretation overlooks structural confounds baked into the experimental design. For policymakers who have invoked tunneling to explain why people in poverty take payday loans or skip subsidized insurance, the message is sobering. The behavioral gaps between ‘rich’ and ‘poor’ players were real, but they arose because the game offered the two groups fundamentally different choices, not because scarcity rewired anyone’s attention. Understanding poverty, the authors conclude, may depend less on probing the minds of the poor and more on mapping the constrained decision spaces they are forced to navigate.

Subject of Research: Re-analysis of the evidence for an attentional tunneling effect caused by resource scarcity

Article Title: Revisiting evidence for an attentional tunneling effect caused by scarcity

Article References: tho Pesch, F., Langenhoff, A. F., & Srinivasan, M. (2026). Revisiting evidence for an attentional tunneling effect caused by scarcity. Nature Communications, 17(1), Article 10233. https://doi.org/10.1038/s41467-026-78074-y

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78074-y

Keywords: scarcity, attention, tunneling hypothesis, poverty, decision making, Angry Blueberries, replication, cognitive load, borrowing, behavioral economics, Nature Communications, re-analysis